Chimp Challenge 2011

the_cultie · May 6, 2011

Well it's that time of the year again, the annual Chimp Challenge has started and we're asking for everyone who has a love for science, doesn't want Team32 to come last (where we are at the moment

) or just loves to fling poo at the other teams, to come help us out.

The info thread is here:
http://www.overclockers.com/forums/showthread.php?t=674871

Go here for the hourly updates on the race:
http://www.overclockers.com/forums/showthread.php?t=675979

The Hardware Canucks team has put together a wonderful stats page that can be viewed here with all the teams points and current production:
http://www.hardwarecanucks.com/chimp_challenge/2011//#

This year the race is being run for 10 days from the 5th to the 15th of May so you know when you can finish up and return to your regular work; Sunday, May 15th at 12pm, Pacific Daylight Time (UTC -7)

If you have any questions about getting started or setting up the client just ask, there's guys there all the time waiting to answer questions, especially if means we beat MaximumPC. :rock:

So please, fold for Team32. Help Overclockers get to the top :salute:

Bobnova · May 6, 2011

I've already switched over till the contest is ending.
It won't set the Rosetta contest back much, and the poor FAH team is getting stomped

the_cultie · May 6, 2011

Thanks Bob. Yes we are getting stomped, it's highly unusual, hopefully we'll pull off our mid way surge.

virtualyoda · May 6, 2011

Got my Zeons, an hd4800 , 8600gt and 2 8800gts on the case, Not much i know, but it all counts,

the_cultie · May 6, 2011

Exactly, it all adds up

What shader clock are you running on the 2 8800GTS? You should be able to get close to 1947MHz if you crank the fan up a little and assuming its the G92 based 8800GTS.

xsuperbgx · May 6, 2011

I will swap over tonight or tomorrow.

drshivas · May 6, 2011

I feel for ya, being a T32 alum and all. But I have no GPUs and blew away my Linux VMs long ago. What's the state of the WinSMP and uni-processor clients nowadays?

Adak · May 6, 2011

We have one client that handles most everything, now. A brand new one will handle everything, but it's not out on the d/l page, yet. Still a few bugs to work out, and monitoring programs need to be changed to work with it.

For Windows, the high performance all in one's are here:
http://folding.stanford.edu/English/DownloadWinOther

For others, remove the "WinOther", part of the above url.

Happily, the FAH clients are easier now -- at last.

muddocktor · May 7, 2011

That's good to hear, Adak. If I get in from the rig in time to help contribute, I'll see if I can switch a system or 2 over.

But let me tell you, the client better be easy to set up because I haven't messed with F@H for some years now.

the_cultie · May 7, 2011

It's quite easy to setup and we have an excellent Super Sticky with all the details. If you're not sure just post a question and we'll have you running in no time.

Adak · May 7, 2011

muddocktor said:
That's good to hear, Adak. If I get in from the rig in time to help contribute, I'll see if I can switch a system or 2 over.

But let me tell you, the client better be easy to set up because I haven't messed with F@H for some years now.

The current FAH clients are stable, and the earlier complexity of setup for multi-core processing, are gone. Also, single and multi-core processing is done on the same client. Multicore is indicated by adding the -smp into the extra parameters, during set up.

You'll remember the setup questions for a new client. These are the special values for the race:

Handle: T32monkeys
passkey: 63db8ce5557ac843556921df5f3a5f95
size of work unit: big

Advanced parameters: y
extra parameter: -smp

Everything else is the same as before.

<< Glad to have you on board! >>

HankB · May 7, 2011

I've temporarily defected from the Rosetta contest (on my main cruncher) and switched my GPU from SETI to FAH as well.

Adak said:
The current FAH clients are stable, and the earlier complexity of setup for multi-core processing, are gone.

I can't speak to the older versions, but the new one was only slightly more difficult to set up than adding a new BOINC project. (I actually had to type a command

) The GPU app was a little trickier, but that's because I'm running Linux and already had a newer version of the required nvidia dev kit installed. But I got that working and found a kernel that provides a good balance between CPU and GPU performance. My rig is happily crunching along about 15-20K ppd I think.

It's a 10 day contest (8 days left) so hopefully Rosetta won't slip too much while I'm away.

muddocktor · May 7, 2011

Adak, what about setting up for both gpu and cpu, both on the same system? Is that possible? Because, since I will be coming in from the rig towards the last part of the challenge (on the 11th), I'm thinking of using my main system to help with. That's a 980X and GTX 570 in that system, with 6 gigs of ram. I could also throw my Sandy Bridge system over temporarily too, which has a 9800 in it, but only 4 gigs of ram.

Adak · May 7, 2011

muddocktor said:
Adak, what about setting up for both gpu and cpu, both on the same system? Is that possible? Because, since I will be coming in from the rig towards the last part of the challenge (on the 11th), I'm thinking of using my main system to help with. That's a 980X and GTX 570 in that system, with 6 gigs of ram. I could also throw my Sandy Bridge system over temporarily too, which has a 9800 in it, but only 4 gigs of ram.

Sure, that's one of the benefits of the newer FAH clients.

For Windows, go to the download page I mentioned, and your browser will put the more suitable client candidates for you, right on the top of the list.

For the CPU's folding, be sure to include the flags I mentioned earlier. With the new bonus system points, the CPU's are even more productive than the GPU's, if you have Hyperthreading on, and Bloomfield or above, at 3.5 GHz or above.

drshivas · May 7, 2011

Adak, which one should I download for just CPU folding? Everything at the top of the page mentions GPU.

torin3 · May 7, 2011

muddocktor said:
Adak, what about setting up for both gpu and cpu, both on the same system? Is that possible? Because, since I will be coming in from the rig towards the last part of the challenge (on the 11th), I'm thinking of using my main system to help with. That's a 980X and GTX 570 in that system, with 6 gigs of ram. I could also throw my Sandy Bridge system over temporarily too, which has a 9800 in it, but only 4 gigs of ram.

If you are willing to run -bigadv work units (take 1 to 2 days to run), you will get almost the same PPD running only CPU as you would running CPU & GPU (due to the GPU client slowing down the CPU client). Running -bigadv WU on a 980x and a Sandybridge should get you about 50-60K ppd per machine. You need HT on, so 12t on the 980x and 8t on the Sandy.

The paramters you need are -smp and -bigadv to get the big WUs, and they can be entered in the advanced settings>extra paramters line.

torin3 · May 7, 2011

drshivas said:
Adak, which one should I download for just CPU folding? Everything at the top of the page mentions GPU.

For your i7 and the Sandy Bridge, the 6.34 windows client is the one you want. 3rd up from the bottom. You need hyperthreading on. Answer yes when it asks for Advanced Settings, and no for Advanced Methods. On the extra paramters line you want -smp and -bigadv.

Adak · May 7, 2011

Go to the high performance Widnows client d/l page:
http://folding.stanford.edu/English/DownloadWinOther

And get the 6.34 version. 64 bit is preferred, if you have the 64 bit version of your OS.

The values to enter that are special include:

Handle: T32monkeys

Team 32

passkey: 63db8ce5557ac843556921df5f3a5f95

size of work unit: big

Advanced settings?: y
extra parameter: -smp

If you want to fold the bigadv work units, then you need an 4 core hyperthreaded and overclocked cpu with hyperthreading on, and 4GB of memory - slightly more preferred. BIG bonus points, but they take 2 days plus to fold one wu, and you have to use the passkey listed above (or fold 10 smp wu's successfully first), to qualify.

If the above is OK with you, then add the -bigadv flag to the extra parameter:
-smp -bigadv

I would not add it as a service if it asks you, because services have no window to watch, and this is a beta client (although a very good one, still...)

And thanks for joining us!!

muddocktor · May 8, 2011

That sounds good guys. I think I will go the bigadv route then, since that yields more points and my systems are capable of it. The 980X is at 5.0 and the SB is at 4.5 and both rock stable there. I can also put my 930 test system on them too, since it's also at 4.0 with 6 gigs of ram and I might also be able to throw the 875k system on too, but I have to check it out first. My daughter told me it powered down about a week or so ago and won't power up, so I don't know what's wrong with it at present. BTW, I run all these systems with HT on already and are stable, so I don't have worries about that.

drshivas · May 8, 2011

with -bigadv there's no more work. I took it off and am using -smp. Glad to see I don't need to install that SMP framework first anymore.

I tried to put this on my C2D T6500 lappy, but with only -smp enabled, it folded on one core only. I killed it figuring it wouldn't make the deadline.

EDIT. NM, my Gulftown picked up work with bigadv. Both my big irons are folding for T32. Those console configs brought back memories.

Oh and FAH crashed my 2600K! I guess this SMP is more stressful than Rosetta and IBT! Had to bump up the vcore a bit.

Chimp Challenge 2011

Member

Senior Member

Member

Registered

Member

Benching Team Leader

Sith Lord

Senior Member

Retired

Member

Senior Member

Member

Retired

Senior Member

Sith Lord

Member

Member

Senior Member

Retired

Sith Lord

Similar threads